expression of Your Favorite Gene (YFG) in yeast David Nelson Mar. 27, 1997
Done
Last modified Mar. 27, 8:30AM
Today we are going to talk about gene expression in yeast. This will include Saccharomyces
cerevisiae and Pichia pastoris used in secreting proteins to the medium. We will talk about
the nuts and bolts of how to express a gene in these systems and then we will cover some real
applications. Before we start, I thought it would be fun to share this example of yeast expression in
science fiction from Issac Asimov who held a Ph.D. in biochemistry from Columbia University
which he earned in 1947.
excerpt from I, Robot (1950)
"In the first place, by far the largest crop we deal with ...is yeast. We have upward
of two thousand strains of yeast in production and new strains are added monthly. ...these strains
of yeast have each their peculiar properties. The beef steak you thought you ate toady was
yeast. The frozen fruit confection you had for dessert was iced yeast. We have filtered yeast juice
with the taste, appearance, and all the food value of milk."
brief review of high copy and low copy vectors
When I talked with you a month ago about yeast genetics, we covered the high
copy (2 micron circle) vectors that have a yeast episomal sequence added to them. This keeps
them expressed at 10-40 copies per cell, for over production of cloned genes. The high copy
vectors can be used in conjunction with any promoters available. A strong promoter on a high
copy vector can produce a lot of transcript and protein. There is a very high copy vector (100-200
copies per cell) with a 2 micron circle and the leu2-d gene as selectable marker. leu2-d has a deletion
of a part of LEU2. It supports growth on leu minus media but not without forcing higher copy
numbers of the vector to compensate for the poor activity of the leu2-d gene product.
The low copy CEN vectors have a yeast centromere and are maintained at 1-3
copies per cell. These can also be used with any promoter.
For strict one copy expression, integrative vectors can be used to place the gene you
want on a yeast chromosome. You may have to show that you only introduced one copy into the
genome to satisfy purists (done with a Southern Blot).
different promoters used in yeast expression
For expression of yeast genes in yeast, to determine the effects of mutations,
it is generally best to use your own gene's promoter in a CEN Plasmid so
expression is similar to the wild type gene. However, there are a variety of
promoters to choose from for various purposes.
1) The Gal 1,10 promoter This promoter is inducible by galactose. It is
frequently valuable to be able to turn expression of your gene on and off so
you can follow the time dependent effects of expression. The Gal promoter
is a little bit leakey, so you do not have complete absence of YFG in the
absence of galactose. The Gal 1 gene and Gal 10 gene are adjacent and
transcribed in opposite directions from the same promoter region. The regulatory
region containing the UAS sequences can be cut out on a DdeI Sau3A fragment
and placed upstream of any other gene to confer galactose inducible expression
and glucose repression.
2,3,4) PGK, GPD and ADH1 promoters. These are high expression constitutive
promoters. PGK = phosphoglycerate kinase, GPD = glyceraldehyde 3 phosphate
dehydrogenase, ADH1 = alcohol dehydrogenase
5) ADH2 promoter. This gene is glucose repressible and it is strongly transcribed
on non-fermentable carbon sources (similar to GAL 1,10 except not inducible by
galactose.
5) CUP1 promoter This is the metalothionein gene promoter. It is activated
by copper or silver ions added to the medium. The CUP1 gene is one of a few
yeast genes that is present in yeast in more than one copy. Depending on the
strain, there can be up to eight copies of this gene.
6) PHO5 promoter. This gene is a secreted gene coding for an acid phosphatase.
It is induced by low or no phosphate in the medium. The phosphatase is secreted
in the chance it will be able to free up some phosphate from the surroundings.
When phosphate is present, no PHO5 message can be found. When it is absent it
is turned on strongly.
steroid inducible expression
Keith Yamamoto's lab developed an inducible system in yeast similar to the ecdysone
system we talked about for mammalian cells. They placed the rat glucocorticoid receptor
gene behind the constitutive GPD promoter to express the rat glucocorticoid receptor in
yeast. A second vector was made with 3 glucocorticoid response elements upstream of
the CYC1 gene minimal promoter. (cytochrome c gene). A cloning site was placed after
this so YFG could be placed under control of the 3GRE/CYC1 promoter. Both vectors
were high copy vectors. This system works well with dose dependent expression when
steroid hormone is added to the medium. Response time is rapid with t1/2 of 7-9 minutes
after addition of hormone.
copper inducible expression
The CUP1 promoter can be used to make a gene inducible by copper or silver ions. Of
course, they only use silver ions at Howard Hughes Labs. We are in the process of
placing the AAC2 gene under CUP1 regulation, which should provide a degree of control
of the level of expression based on the amount of copper in the medium. Copper is toxic
and your strain should be tested to see how well it tolerates copper before making a CUP1
construct.
heat shock expression
By placing the UAS from a heat shock gene in front of the minimal CYC1 promoter, you
can place YFG under heat shock induction. This is a specialized requirement ususally
used in studies of heat shock response. It is unlikely the average researcher would use this
method.
Other Considerations
In addition to promoters, cellular location is an important consideration. To
get YFG expressed in the right location you may have to add a targeting
signal.
1) Steroid receptor fusions. Steroid receptor vectors have been developed
that cause targeting to the nucleus upon addition of steroid hormones. These
sequences are fused in frame with YFG so it gets carried along with the
steroid receptor.
2) Mitochondrial targeting signals. These are rather non-descript signals.
Almost any amphiphilic helix with positive charges and no negative charges
can serve as a signal. In fact, an experiment was done that fused random
DNA to a reporter gene and about 25% of the products could be targeted to
mitochondria.
3) Peroxisome targeting signal The C-terminal sequence SKL targets to
peroxisomes. More detailed work has shown some mutations are tolerated
in this sequence, so the SKL sequence is not the only acceptable one. This method
may be more important for Pichia expression systems since peroxisomes make up the
majority of the cell volume when Pichia is grown on methanol.
4) Nuclear targeting signals. The sequence from the N-terminal of the large
T antigen of SV40 virus targets to the nucleus. This is frequently used in
selection schemes to isolate nuclear pore components.
5) Secretion is also an option. With the right leader sequence, YFG can be
secreted into the medium. In S. cerevisiae, few proteins are secreted. These
include invertase, mating factor alpha and PHO5 and SUC2. Signal sequences from
any of these genes can direct YFG to the secretory pathway in yeast.
Secretion has been exploited more extensively in a kit that uses the
yeast Pichia pastoris to secrete a rather high concentration of protein into the
medium. Grams of protein per liter of media have been generated, and it is
nearly pure protein, because yeast do not secrete that many proteins. The
vectors used express YFG with a leader sequence that is precisely trimmed
upon export, so the final product is just the native protein without the leader
sequence. There may be some sequence modification at the fusion joint.
Pichia pastoris
Pichia is similar to S. cerevisiae in practice, but it can give 10 -100 fold higher levels of
expression for a foreign gene. Pichia is a methylotrophic yeast that can use methanol as
sole carbon source. It does this by formation of formaldehyde and hydrogen peroxide
inside peroxisomes. The enzyme that carries out this reaction is alcohol oxidase. This
enzyme is made in large amounts in peroxisomes. Pichia expression vectors use the
AOX1 promoter to drive expression of foreign genes. This gene can be induced by
methanol so that AOX1p is about 30% of soluble protein in the cells. Glucose represses
the AOX1 gene. Not even a trace of it can be seen in the presence of glucose. For
expression of YFG, grow Pichia on glycerol to derepress the gene. Even then, no message
is made unless methanol is added. (see Faber et al. Review: methylotrophic yeasts as
factories for the production of foreign proteins. Yeast 11, 1331-1344 1995)
Mutants in AOX1 are still able to grow on methanol, but they grow slowly. This is called
the MutS phenotype. Mut+ cells are wild type for AOX1 and grow well on methanol.
Glycosylation in Pichia is different from S. cerevisiae. 8-14 mannose are added per chain
compared to 50-150 in S. cerevisiae, and the terminal linkages in S. cerevisiae are alpha
1,3 glycan linkages. This is not true in Pichia. Pichia may be more like higher eukaryotes
in its glycosylations. S. cerevisiae expressed proteins are hyper antigenic and not suitable
for therapeutic use.
All Pichia vectors are integrative vectors with HIS4 as selectable marker. I did not find
out if this is due to lack of knowledge about the ARS sequence or some other reason. It
did not seem to be necessary to integrate each time if the plasmids could be maintained in
the cells. The site of integration can be directed to his4 or AOX1 depending where the
plasmid is cut. The plasmid will integrate in the gene that has the ends on the linearized
plasmid. A cut in HIS4 will target to his4. A cut in AOX1 will target to AOX1.
Integration at AOX1 can result in replacement of the the AOX1 gene with your gene.
This will cause the cells to be MutS and grow slowly on methanol. Integration at his4 will
leave the cells as Mut+. Both types of cells are desirable becuse you don't know in
advance which will produce more of your protein.
Once you have colonies of MutS and Mut+ phenotype grow the cells and induce with
methanol then assay for your protein or check for it on SDS PAGE. Follow standard
procedures given in the kit to optimize expression (either intracellular or secreted).
Adrenocortical Yeast
When we talked about the Blue Rose Project, we encountered the idea of gene expression
for the sake of engineering novel pathways into an organism. The concept of gene therapy is
based on restoring defective pathways to cure a genetic disease. Thus, pathway engineering must
be considered of at least equal importance to overproduction of a protein for the sake of purification
and study of that protein. On Monday, we will talk about an example of pathway engineering in
yeast rather than roses. This is a powerful idea that is being commercially exploited by multiple
companies. Today I have two more examples of pathway engineering in yeast that illustrate the
economic importance of this technology.
The pharmaceutical industry has traditionally manufactured drugs by chemical synthesis.
This frequently involves many steps with overall low yields. Competitors are always on the
lookout to reduce the number of steps in an important synthesis, or to improve yield. Often one
consideration is the cost of treating wastes produced in the process, especially if they contain toxic
chemicals or heavy metals. I heard a talk about a year and a half ago on the case of industrial
manufature of steroid hormones, specifically hydrocortisone. The speaker was R. Spagnoli of the
company Roussel Uclaf. He outlined the history of steroid manufacture, with an account of the
earliest synthesis that took about 40 steps. This was slowly improved upon by more sophisticated
chemical strategies to a much smaller number of steps. He then introduced the concept of
bioconversions, or getting microorganisms to do some of the steps previously done by chemists.
In the end the modern manufacture method now requires 8 steps, including some bioconversions.
Dr. Spagnoli's goal was to engineer in yeast the pathway for direct biosynthesis of hydrocortisone
from cholesterol. In mammals this is done in the adrenocortex by five enzymes in two different
compartments, the ER and the mitochondria. Four of the five enzymes are cytochrome P450s.
The pathway starts in the mitochondria with the cleavage of the lipid side chain to make
pregnenolone. This then moves to the ER where it is oxidized and hydroxylated by three more
enzymes to make 11-deoxycortisol. This moves back to the mitochondria to be converted to
hydrocortisone. To engineer this pathway into yeast would require expression of five enzymes in
the correct compartments and adrenodoxin and adrenodoxin reductase needed in the mitochondria
for electron transfer to the P450s. Also, yeast does not make cholesterol. It makes ergosterol
instead and it does not take up cholesterol from the medium, so a way has to be found to get
cholesterol into the yeast. This was a very ambitious project, but the goal would be biosynthesis
of a valuable steroid in one bioconversion step, with no waste products except yeast.
For this process, the research team started at the last step and worked backwards. As of
the talk, they had expressed the last two P450s, one human microsomal enzyme in the ER and one
bovine enzyme in the mitochondria along with the two electron carriers (both bovine) in the
mitochondria. The engineered system could successfully convert 17 hydroxy progesterone to
hydrocortisone. This meant they had reconstituted the last two steps of the pathway. This is an
incredible feat and the rest will surely follow (and it may be done by now). They have already
made four mammalian proteins in this yeast simultaneously and with correct targeting. These were
all expressed off a single vector. It should not be any more difficult to do the last three proteins.
Polyester Yeast
In the April 1997 Scientific American, a two page add (pp.18-19) from Dupont asks the
question in big bold letters: Use yeast to turn sugar into other molecules? Then they go on to tell
you they are not talking about alcohol, but about a polymer called polytrimethylene terephthalate
(3GT). This polymer is more versatile than traditional polyester abbreviated (2GT) that is made
from ethylene glycol (2G) [1,2 ethandiol, HO-CH2-CH2-OH]. The process of making the two
polymers is similar but the monomers that go into each are different. One factor preventing the
commercial production of 3GT is the cost of one of its monomers, trimethylene glycol (3G) [1,3
propanediol, HO-CH2-CH2-CH2-OH]. 3G is made by some bacteria starting from glycerol. Some
naturally occurring yeasts can make glycerol from sugar. However, no organism known does both.
So in comes pathway engineering. Dupont joined forces with a company called Genencor
International to add the bacterial genes for conversion of glycerol to 3G into yeast. They have done
it, and the process involves no heavy metals, petroleum or toxic chemicals.
The carbon source for the process is glucose.